hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jim Kellerman (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-1124) Balancer kicks in way too early
Date Tue, 13 Jan 2009 16:13:12 GMT

    [ https://issues.apache.org/jira/browse/HBASE-1124?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12663370#action_12663370
] 

Jim Kellerman commented on HBASE-1124:
--------------------------------------

Ron-En is correct. The region balancer does not kick in until all regions have been assigned.
See master.RegionManager.assignRegions lines 187 - 204.

HBASE-1104 should have had no effect on region assignment other than preventing assignment
to multiple servers.

During any region assignment, the master will assign up to RegionManager.maxAssignInOneGo
(default = 10). Depending on the number of regions and the responsiveness of the region servers,
some may get (a lot) more regions assigned than others. After the initial assignment of all
regions, region balancing then kicks in and may result a lot of churn until it has balanced
the region load.

> Balancer kicks in way too early
> -------------------------------
>
>                 Key: HBASE-1124
>                 URL: https://issues.apache.org/jira/browse/HBASE-1124
>             Project: Hadoop HBase
>          Issue Type: Bug
>            Reporter: Andrew Purtell
>             Fix For: 0.19.0
>
>
> Balancer kicks in before all regions are assigned out. Causes confusion. Master won't
accept OPENs from "overloaded" HRS. Master is slow to respond to UI and HRS during. Master
sometimes takes too long to respond to a HRS heartbeat and so the HRS will reinit. This causes
more confusion.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message